[WIP] Monitoring Interpersonal Relationships

Third blog post (hopefully final) to continue my delusion of correlating interpersonal relationships to design patterns through an SRE view.

On conjecture that all relationships will fail, we need to break out parts of the system’s inevitable failure and see where we can improve.

Following my last post’s “Normal Accident” theory applied to relationships, let’s see where we can proactively monitor and apply system design principles to.

The antithesis of tight coupling is loose coupling. If two individuals are so intertwined that one’s volatility immediately triggers the other’s collapse, the system is operating with zero margin for error. This is a design flaw. Naturally, we need error budgets. We cannot expect a system to have 0 errors. If you choose to love a partner with a known “bug” (EX:short fuse), you have accepted that sub-system into your architecture. You have to budget for the 5% of mornings where the “asshole” sub-system runs by default. You don’t fix the person. You build a buffer that accounts for them so that they on their own want to improve. #selfhealing

How can you aleviate interactive complexity so that one failing subsystem doesn’t cause a cascading failure? Why through direct indicators.

Now for key metrics: When the error-prone sub-system runs, how long does it take for the relationship to return to its baseline state? Calculate the MTTR for gauging reseliency.